Research on evaluating water pollution determinants using multiple logistic regression

Faulty of Arts and Science, University of Toronto, Toronto, M5S 1A1, Canada

^* Corresponding author: Ruoyan.shi@mail.utoronto.ca

Abstract

Water pollution is a pivotal challenge, underpinning urgent conversations around environmental sustainability, public health, and ecosystem viability. This research aims to assess the degree of water pollution, dissect and understand the myriad factors contributing to it, and pave the way for formulating effective mitigation strategies and policies to preserve the integrity of water bodies worldwide. It highlights that rapid industrialization, population growth, and agriculture cause pollution. Industrial activities release pollutants like heavy metals, while agriculture contributes through runoff. Urbanization also exacerbates the problem. The study uses a dataset from Kaggle and selects variables like aluminium, ammonia, etc. A multiple logistic regression model analyses factors affecting water potability. Results show that aluminium, chloramine, and ammonia positively correlate with potability, while uranium and barium have negative ones. Interaction terms added to the model improve its fit. The study emphasizes understanding individual contaminants and their interactions for effective water management strategies. Accounting for these interactions enables a more comprehensive understanding of the factors affecting water safety. These insights are crucial for developing targeted and effective water management strategies that ensure safe drinking water and support public health.